Why hold several factors rather than the best one?
If momentum has the highest paper premium, hold momentum. The previous lesson is the reply: the highest premium came with the fastest crash, and the second-highest with the longest drought. Holding several factors is the portfolio-theory course's free lunch applied one level up — diversification across premia rather than across shares.
The correlations that make it work
Value and momentum are the pair that matters. A value sort buys what has fallen; a momentum sort buys what has risen; over most windows their long-short returns have been negatively correlated, so a year in which value loses is more often than not a year in which momentum gains. 2017 to 2020 was a momentum decade and a value drought; 2022 reversed both. Quality sits between them and has tended to hold up when value falls. Low volatility hedges the market factor itself.
A portfolio holding two factors with a negative correlation earns the average of their premia with much less than the average of their drawdowns — the correlation arithmetic from the portfolio-theory course, with factors as the assets.
Two ways to combine
Side by side: hold a value fund and a momentum fund at fixed weights and rebalance between them. Simple, transparent, and each fund pays its own turnover. Integrated: one index scores every share on several factors at once and holds the names that score well on the combination. The integrated version avoids buying a name in one fund and selling it in another, which is real money at momentum's turnover; it also makes the exposures harder to read from the holdings, because no name is there for one reason.
Either way, the result is a portfolio whose factor loadings — the exposure lesson's numbers — are chosen rather than inherited.
What combining does not do
It does not remove the market. A long-only multi-factor fund is still a fund of shares with a beta near one, and in 2008 every factor fund fell with the market, because the market factor is the largest exposure any of them carry. Combining factors diversifies the tilt; the allocation between shares and bonds diversifies the rest, and that is the allocation course's job, not this one's.
Nor does it raise the premium. A blend earns a blend, and the ten-year figures measured on 2026-09-04 — momentum 332%, value 231%, quality 280%, low volatility 164%, against 316% for the market — average to about 252%, which is below the market itself. The case for combining is the path, not the destination: fewer years like 2017 to 2020, and no quarters like spring 2009.
In the data
A blend's exposures are the weighted average of its funds', and each fund's tilt is readable from its holdings. The first check on whether two funds diversify each other or double up is the overlap at the top. The momentum fund and the value fund:
On 2026-09-04 the two held 125 and 308 names; ten each are shown here, and the overlap on the day is the exercise below.
Try it now
Count the overlap between the two top-ten lists above. Zero or one is the diversification working; five is a sign that one factor has become the other for now.
The market exposure that no blend of factor funds removes, at the range where 2008 is visible:
Measure October 2007 to March 2009. Every factor fund fell by roughly that; write the sentence that says what a factor blend is for, and what it is not for. 3. Reopen Is holding ten stocks the same as holding one stock ten times? and say why the same distinction applies to holding two value funds versus a value fund and a momentum fund.